Building Redundant Switched Topologies

Question 1

What is STP (in general) and what is it’s main goal?

Answer 1

Spanning Tree Protocol

It is used to prevent loops for forming when creating redundant links on the network.

NOTE: If a loop forms, then things such as a broadcast storm can start and end up taking down the network.

Question 2

What does STP use for determining if there are redundant connections and how does it handle those redundant connections?

Answer 2

BPDU (Bridge Protocol Data Units)

Devices using STP send out BPDU’s out it’s ports to determine the links on the network and it will disable redundant links. If a primary link goes down, it will notice this using BPDU’s and enable the previously disabled redundant link.

Question 3

How does STP find the best path?

Answer 3

1. Elect the root bridge
   - it becomes the center of the network in terms of STP (logical network)
2. Find the best path to the root
   - lowest cost
   - lowest bridge ID (if the costs are the same)
   - lowest port number (if the cost and bridge IDs are the same)
3. Block whatever is left over

Question 4

What are 3 conditions that can occur in a redundant switch topology if no protocols are monitoring link forwarding state?

Answer 4

1. Continuous frame duplication - also called a broadcast storm. Switches flood broadcast, multicast and unknown unicast frames to all ports except the originating port endlessly.
2. Multiple frame transmission - multiple copies of the same multicast frame could cause unrecoverable errors in protocols that don’t expect the same copy of a frame.
3. MAC database instability - different ports on the switch receive copies of the same frame, causing the switch to work harder to process this in the MAC database, thus consuming it’s resources and could lead to impairing it’s ability to forward other normal traffic.

Question 5

What is the overall process of STP?

Answer 5

1. All interfaces on all switches in the spanning tree topology start in blocking mode.
2. The switches elect a root bridge. - Based on lowest BID, if they have the same bridge priority then the lowest MAC wins. If the root bridge switch fails, then next lowest BID takes over as root bridge. Cisco enables the configuration of a root bridge and secondary root bridge.
3. Each nonroot bridge determines a root port. It is based on the best path to the root bridge. If there are two best paths, the path with the lowest Bridge ID is selected. If the lowest Bridge ID is the same, then the lowest port on lowest Bridge ID side of the connection is the best path to the root bridge.
4. A designated port is selected for each segment. This is based on the lowest root path cost. The designated port on a segment is on the switch with the lowest root path cost. If they have the same root path cost, then the lowest Bridge ID side of the segment is the designated port. On root bridges, all switch ports are designated ports. Each network segment will have one designated port.
5. The root ports and designated ports transition to the forwarding state and any other ports (nondesignated ports) stay in the blocking state.

Question 6

For STP, what are the costs of each type of link?

Answer 6

Question 7

What are the different STP Port Roles?

Answer 7

Question 8

What are the different RSTP port states?

Answer 8

Question 9

What is the IEEE standard for Common Spanning Tree (CST) and how long does it take for a CST port to come online fully?

Answer 9

IEEE standard is 802.1D

It was developed in 1993. It takes 30 - 50 seconds to determine the root bridge, etc. and bring ports/communication online.

Question 10

What is PVST+ and what did it do?

Answer 10

Per-VLAN Spanning Tree

Cisco proprietary solution that allowed you to setup different Root Bridges, one for each VLAN. You can end up blocking different links and enable some forms of load balancing.

Requires more CPU and memory resources than common spanning tree (802.1D).

Question 11

What is RSTP and the IEEE standard for it?

Answer 11

Rapid Spanning Tree - 802.1w

It brings the time for a new connection to come online down to ~2 seconds. Requires more resources than original STP, but less that PVST+.

Faster convergence time by redefining port roles and BPDU exchanges. However, it’s not per vlan, so it does not address the suboptimal traffic flows (no load balancing like PVST+ and Rapid PVST+)

Question 12

What is Rapid PVST+ and what are the pros and cons of it’s use?

Answer 12

It’s Cisco’s proprietary per-VLAN implementation of the RSTP (802.1w) standard.

This means it does provide faster convergence times (<2 seconds) and also allows per-VLAN for load balancing the used connections.

However, it does use the most CPU & memory resources out of the different spanning tree protocols.

Question 13

What is MSTP and it’s IEEE standard?

Answer 13

Multi-VLAN Spanning Tree Protocol - 802.1s

It was inspired by Cisco’s proprietary Multi-Instance STP (MISTP).

It allows for multiple VLANs to be grouped into a spanning tree instance (RSTP instance). So it has all of the benefits of Cisco’s Rapid PVST+, but not require as many CPU and memory resources.

Question 14

Cisco Catalyst switches, by default, use which STP protocol?

Answer 14

PVST+

Enabled on all ports in VLAN1

Resulting in a slower convergence after topology change than with RSTP.

NOTE: The STP protocol can also be dictated based on how old your switch is and what manufacturer.

Question 15

How is the STP root bridge elected?

Answer 15

By default it will be elected based on the lowest Bridge ID.

Bridge ID = Bridge Priority + Bridge MAC

Bridge Priority default = 32768 (for industry standardized switches)

NOTE: When you run PVST+ or Rapid PVST+, then you take the default 32768 + vlan# = Bridge Priority

Question 16

What are the stages that STP & PVST+ go through to ensure loops are not created?

Answer 16

1. Blocking - For up to 20 seconds, the port remains in the blocking state.
2. Listening - For 15 seconds, the port listens to BPDUs that it received and listens for new topology information. If a better BPDU is received the port transitions back to blocking. If not, it transitions to learning state. In listening state, the port does not populate the MAC address table with the addresses it learns and it does not forward any frames.
3. Learning - For up to 15 seconds, the port updates the MAC address forwarding table, but it does not begin forwarding.
4. Forwarding - Once the switchport is certain of no loops forming, it enters the forwarding state. It still monitors for topology changes that could require it to transition back to the blocking state to prevent a loop.

Question 17

What are the two features that Cisco designed as enhancements to STP to limit the amount of time for access devices to reach the forwarding state?

Answer 17

1. PortFast - when enabled on an access port, the port bypasses the STP listening and learning states to move straight to forwarding.
2. BPDU guard - puts port in ‘errdisabled’ if the port receives any BPDUs.

Question 18

How is PortFast implemented in RSTP?

Answer 18

• PortFast is known as an edge port concept, instead of PortFast
• All ports are assumed to be edge ports until a BPDU is received, then it loses it’s edge port status and becomes a normal spanning-tree port.

Question 19

In what port state will a port accept a BPDU while using RSTP?

Answer 19

A port will accept and process BPDU frames in all port states.